The process when carried out in this manner has one defect, for it is necessary to allow the precipitate to settle in order to see clearly whether an additional drop of the silver solution will produce further precipitation, or whether it will merely cloud the fluid; this defect can, however, be remedied by means of a so-called “indicator.” A few drops of a concentrated solution of neutral yellow potassium chromate should be added to the fluid to be examined, which is thereby coloured yellow; the solution is then shaken or stirred with a glass rod, while the silver nitrate is dropped in. Every drop of the silver solution will cause a red precipitate, the colour of which however disappears on stirring so long as there is any chlorine present; only when the silver solution has precipitated all the chlorine does the red colour become permanent, and thus the change of colour of the whole fluid from yellow to red shows with exactness the complete precipitation of the chlorine. For practical purposes all that is required is the so-called decinormal silver solution, and from the number of cubic centimetres of this solution which are required to precipitate all the chlorine the total amount of chlorine present can be readily calculated.

In steeping smaller objects before examination the whole of the water should be well stirred with a glass rod or poured two or three times from one vessel into another: 100, 50 or 25 cubic centimetres are then poured into a graduated glass or drawn up into a pipette. The water should be drawn up by suction slightly above the level of the mark upon the stem of the pipette, the upper end of which is immediately closed with the thumb. By slightly raising the thumb the water is allowed to run off until its upper surface is exactly level with the mark. The amount taken is then placed into a beaker (for 100 cubic centimetres a beaker of 400 c.c. capacity should be used), and, after the addition of a few drops of a solution of potassium chromate, is examined by titration. In the treatment of large objects, for which tubs are required, the necessary quantity of water may be drawn by means of a long pipette from the bottom of the tub, where the quantity of salt is always greatest, or through the tap at the bottom of the tub (as was done with the blocks from the Meten Chamber, in which case about 1 litre was drawn off into a glass out of which 100 c.c. were taken for titration). To obtain results which are comparable, care must be taken that the object is always as nearly as possible in the same quantity of water. After placing the larger blocks in the water, one examination should be made during the first few days, when the titration may require 20 c.c. or more of silver solution. There is no need to examine for chlorine while the water is being frequently changed: indeed, in order to economise the silver solution, this need not be begun until the second month, when the water is changed every fortnight.

As has been stated above, it is only necessary to read off the number of cubic centimetres of the solution used in the titration, for the decrease in these figures is a sufficient indication of the progress of the operation, while the diminution of the chlorine-content may be taken as an indication of the simultaneous removal of the sulphates[88]. In the treatment of small objects in distilled water, the process may be regarded as complete if the red colour is obtained on the addition of from one to two drops (i.e. about 1 ⁄10 to 1⁄ 5 cubic centimetre) of the silver solution. If, when tap-water is used, and is being changed at intervals of a fortnight or a month, the estimations give a constant result between 0·6 and 1·0, the treatment need not be carried further.

The accompanying table shows the figures obtained from three large blocks from the Meten Chamber. They represent the number of cubic centimetres of decinormal silver solution used for 100 c.c. of the water, which was changed every fortnight. The first column on the left shows the dates upon which the stones were placed in the tubs:

3 Feb.
1890		7 Apr.4 May2 June25 Aug.22 Sep.4 Nov.1 Dec.12 Jan.
1891
3·02·52·31·41·00·90·80·8
25 Apr.
1893		9 May18 July1 Aug.25 Nov.23 Dec.16 Feb.
1894		29 Aug.12 Sep.
6·05·64·92·01·71·50·80·8
29 July
1893		28 Oct.25 Nov.23 Dec.20 Jan.
1894		16 Feb.1 Aug.29 Aug.12 Sep.
3·02·01·51·31·10·70·80·7

When repeated examinations gave a fairly constant result of 0·7-0·8 cubic centimetre the process was regarded as complete, for Berlin tap-water itself contains small quantities of chlorine compounds, 100 c.c. requiring from 0·4 to 0·6 c.c. decinormal silver solution. Before using the water from a well or from waterworks, it should be examined to ascertain the number of cubic centimetres of silver solution required to produce the red colouration. As the amount of chlorine compounds in the water may vary it is advisable to repeat the examination[89].

The following table shows the rapidity with which salts can be completely extracted from small pieces of limestone. The limestones were placed in tap-water in three glass cylinders, each containing 2 litres; the amount of silver solution required for the water was 0·45 c.c. per 100 c.c.

No. of days of soaking in water111224
1Weight in grammes 1073·4 c.c.1·41·10·70·50·5cubic centimetres of decinormal silver solution.
2Weight in grammes 663·0 c.c.1·20·70·5—0·5cubic centimetres of decinormal silver solution.
3Weight in grammes 473·6 c.c.0·80·60·5—0·5cubic centimetres of decinormal silver solution.

These figures show the numbers of the c.c. of silver solution used for every 100 c.c. of the water, which was changed after 1, 2, etc. days as shown above. After 9 to 11 days therefore the stones could be declared free from salt.

If the accuracy of the titration method be considered unnecessary, either on account of the small number or size of the objects to be treated, or for reasons of expense (the outlay required is however very small), a solution of unknown strength may be used. A comparison between the degree of turbidity produced on mixing the silver nitrate solution with the tap-water, and that produced with the wash-water, will enable the progress of the operation to be gauged.